Extending HIRS High Cloud Trends with MODIS

Donald P. Wylie
Richard Frey
Hong Zhang
W. Paul Menzel

12 year trends
Effects of orbit drift and ancillary Tsfc
Comparison with MODIS

July 2002
Cloud Properties from CO2 Slicing

RTE for cloudy conditions indicates dependence of cloud forcing

(observed minus clear sky radiance) on cloud amount () and
cloud top pressure (pc)

pc
(I - Iclr) =    dB .
ps

Higher colder cloud or greater cloud amount produces greater cloud

forcing; dense low cloud can be confused for high thin cloud. Two
unknowns require two equations.

pc can be inferred from radiance measurements in two spectral bands

where cloud emissivity is the same.  is derived from the infrared
window, once p is known.
Different ratios
reveal cloud
properties
at different levels

hi - 14.2/13.9
mid - 13.9/13.6
low - 13.6/13.3

Meas Calc
pc
(I1-I1clr) 1  1 dB1
ps
----------- = ----------------
pc
(I2-I2clr) 2  2 dB2
ps
Generating HIRS Clear Sky Radiances in Cloudy FOVs
Use IR Window Moisture Corrected Brightness Temperature Test
against a priori surface temperature to identify nearby clear sky FOVs
BT11 + aPW * (BT11 - BT12) - Sfc Temp < 2 C
aPW of 0.8 has been used
Sfc Temp estimated from GDAS
Estimate Iclr by interpolating nearby clear FOVs
cld = x x x o o = clr
x x x o o
o x x o o
o x x x o
o x x o o
x x x o o

Attempt to derive CO2 cloud properties in x (note that CO2 cloud

algorithm attempt on x can change FOV to o)
Determining Cloud Presence and Properties with HIRS

Use bands where (I - Iclr) > 1 mW/m2/ster/cm-1 in CO2

slicing estimation of pc

Estimate IRW using IRW radiances

If more than one pc is estimated, use RTE for all bands to

select best one

If no bands qualify, try IR window estimate for opaque cld

If too low in atmosphere, declare FOV clear

GOES Sounder detecting diurnal change of cloud cover
D iu r n a l C h a n g e o f F r e q u e n c y o f C lo u d s d u r in g W in te r 1 9 9 9
8 0 .0
7 0 .0
6 0 .0
5 0 .0
Frequencyof

Occurrence
4 0 .0 C lo u d y ( % )
3 0 .0 H ig h ( % )
2 0 .0
1 0 .0
0 .0
1 :3 0 4 :3 0 7 :3 0 1 0 :3 0 1 3 :3 0 1 6 :3 0 1 9 :3 0 2 2 :3 0
LST

D iu r n a l C h a n g e o f F r e q u e n c y o f C lo u d s fo r S u m m e r 1 9 9 9
8 0 .0
7 0 .0
6 0 .0
5 0 .0 C lo u d y ( % )
Frequencyof

Occurrence

4 0 .0 H ig h ( % )
3 0 .0
2 0 .0
1 0 .0
0 .0
1 :3 0 4 :3 0 7 :3 0 1 0 :3 0 1 3 :3 0 1 6 :3 0 1 9 :3 0 2 2 :3 0

LST
Wielicki et al (2002) CERES deviation of reflected
shortwave flux wrt 1985-89 mean for 20N-20S
NH (36% land)

Tropics (11% land)

SH (very little land)

Determining Cloud Presence and Properties with MODIS

Use MODIS Cloud Mask to determine cloud presence

Calculate Iclr from GDAS

Attempt CO2 slicing estimation of pc on 5x5 FOV average

when (I - Iclr) > 1 mW/m2/ster/cm-1

Estimate IRW using IRW radiances

If no bands qualify, try IR window estimate for opaque cld

Differences in MODIS and HIRS Cloud Property Processing

MODIS HIRS

5 km 20 km

multi-detector single detector

contiguous every 3rd element every 3rd line

uses MODIS cloud mask uses split window comparison

with Tsfc

forward calc of I(clear) interpolate neighboring I(clear)

no radiance bias correction radiance bias correction

MODIS and HIRS global mean CO2 band brightness temperatures
Conclusions

(a) Trends are beginning to emerge in HIRS data; orbit drift

issues; pathfinder reprocessing enabling new look
(b) HIRS & MODIS total cloud cover is roughly the same
over water
(c) MODIS has more high and middle clouds than HIRS over
both land and water surface;
(d) HIRS found more high thin clouds than MODIS in tropics
over both land and water for day and night, but MODIS has
more high thick clouds than HIRS in both tropics & 20-60N.
(e) More work remains